A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images

Zhiwen Wei; Sungjoon Park; Jaeil Kim

doi:10.1007/978-3-031-16919-9_17

A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images

Zhiwen Wei
, Sungjoon Park
, Jaeil Kim

Kyungpook National University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Recently, self-supervised learning(SSL) has shown its great potential in representation learning and been applied to various computer vision tasks. With the success of SSL, which showed performance improvement in natural images, SSL research is actively being conducted in medical image analysis. In this paper, we present a triplet network for the medical image representation learning to learn robust patterns of medical images against global and local changes by comparing latent feature distance between positive and negative pairs with anchors. This approach does not require large batches or the asymmetry of the network. It has been experimentally shown that the proposed method can outperform ImageNet pretrained models and the state-of-the-art SSL methods.

Original language	English
Title of host publication	Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings
Editors	Islem Rekik, Ehsan Adeli, Sang Hyun Park, Celia Cintas
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	181-190
Number of pages	10
ISBN (Print)	9783031169182
DOIs	https://doi.org/10.1007/978-3-031-16919-9_17
State	Published - 2022
Event	5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Virtual, Online Duration: 22 Sep 2022 → 22 Sep 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13564 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
City	Virtual, Online
Period	22/09/22 → 22/09/22

Keywords

Chest X-Ray
Medical Image Classification
Self-supervised Learning
Triplet Margin Loss
Triplet Network

Access to Document

10.1007/978-3-031-16919-9_17

Cite this

Wei, Z., Park, S., & Kim, J. (2022). A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images. In I. Rekik, E. Adeli, S. H. Park, & C. Cintas (Eds.), Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings (pp. 181-190). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13564 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-16919-9_17

Wei, Zhiwen ; Park, Sungjoon ; Kim, Jaeil. / A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images. Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings. editor / Islem Rekik ; Ehsan Adeli ; Sang Hyun Park ; Celia Cintas. Springer Science and Business Media Deutschland GmbH, 2022. pp. 181-190 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{20619efc057a4faea4d89ed60344388a,

title = "A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images",

abstract = "Recently, self-supervised learning(SSL) has shown its great potential in representation learning and been applied to various computer vision tasks. With the success of SSL, which showed performance improvement in natural images, SSL research is actively being conducted in medical image analysis. In this paper, we present a triplet network for the medical image representation learning to learn robust patterns of medical images against global and local changes by comparing latent feature distance between positive and negative pairs with anchors. This approach does not require large batches or the asymmetry of the network. It has been experimentally shown that the proposed method can outperform ImageNet pretrained models and the state-of-the-art SSL methods.",

keywords = "Chest X-Ray, Medical Image Classification, Self-supervised Learning, Triplet Margin Loss, Triplet Network",

author = "Zhiwen Wei and Sungjoon Park and Jaeil Kim",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 ; Conference date: 22-09-2022 Through 22-09-2022",

year = "2022",

doi = "10.1007/978-3-031-16919-9\_17",

language = "English",

isbn = "9783031169182",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "181--190",

editor = "Islem Rekik and Ehsan Adeli and Park, \{Sang Hyun\} and Celia Cintas",

booktitle = "Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings",

address = "Germany",

}

Wei, Z, Park, S & Kim, J 2022, A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images. in I Rekik, E Adeli, SH Park & C Cintas (eds), Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13564 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 181-190, 5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, Virtual, Online, 22/09/22. https://doi.org/10.1007/978-3-031-16919-9_17

A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images. / Wei, Zhiwen; Park, Sungjoon; Kim, Jaeil.
Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings. ed. / Islem Rekik; Ehsan Adeli; Sang Hyun Park; Celia Cintas. Springer Science and Business Media Deutschland GmbH, 2022. p. 181-190 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13564 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images

AU - Wei, Zhiwen

AU - Park, Sungjoon

AU - Kim, Jaeil

PY - 2022

Y1 - 2022

N2 - Recently, self-supervised learning(SSL) has shown its great potential in representation learning and been applied to various computer vision tasks. With the success of SSL, which showed performance improvement in natural images, SSL research is actively being conducted in medical image analysis. In this paper, we present a triplet network for the medical image representation learning to learn robust patterns of medical images against global and local changes by comparing latent feature distance between positive and negative pairs with anchors. This approach does not require large batches or the asymmetry of the network. It has been experimentally shown that the proposed method can outperform ImageNet pretrained models and the state-of-the-art SSL methods.

AB - Recently, self-supervised learning(SSL) has shown its great potential in representation learning and been applied to various computer vision tasks. With the success of SSL, which showed performance improvement in natural images, SSL research is actively being conducted in medical image analysis. In this paper, we present a triplet network for the medical image representation learning to learn robust patterns of medical images against global and local changes by comparing latent feature distance between positive and negative pairs with anchors. This approach does not require large batches or the asymmetry of the network. It has been experimentally shown that the proposed method can outperform ImageNet pretrained models and the state-of-the-art SSL methods.

KW - Chest X-Ray

KW - Medical Image Classification

KW - Self-supervised Learning

KW - Triplet Margin Loss

KW - Triplet Network

UR - https://www.scopus.com/pages/publications/85140482114

U2 - 10.1007/978-3-031-16919-9_17

DO - 10.1007/978-3-031-16919-9_17

M3 - Conference contribution

AN - SCOPUS:85140482114

SN - 9783031169182

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 181

EP - 190

BT - Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings

A2 - Rekik, Islem

A2 - Adeli, Ehsan

A2 - Park, Sang Hyun

A2 - Cintas, Celia

PB - Springer Science and Business Media Deutschland GmbH

T2 - 5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

Y2 - 22 September 2022 through 22 September 2022

ER -

Wei Z, Park S, Kim J. A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images. In Rekik I, Adeli E, Park SH, Cintas C, editors, Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 181-190. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-16919-9_17

A Triplet Contrast Learning of Global and Local Representations for Unannotated Medical Images

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this